Visualizations of Stack Overflow Survey Data¶
By Michael A. Madaio, for the United Nations University, Institute on Computing and Society
August, 2018
Overview¶
Online developer communities boast millions of users - over 29 million on GitHub and over 8 million on Stack Overflow, in 2018. Participation in these communities is becoming one of the primary ways software developers learn new programming languages, improve their skills, develop collaborative projects, and find new job opportunities. [1] [2] [3]
Developers on these sites may ask and answer coding questions to improve their skills (e.g. Stack Overflow), use those skills to contribute to open-source code (e.g. GitHub) and participate in coding challenges (e.g. HackerRank). These platforms are becoming increasingly important to hiring decisions, as recruiters look at GitHub contributions or reputation on Stack Overflow as indicators of developers' skill.
However, despite the promise for these online developer communities to support software developers in their professional development, there are indicators that there may be serious difference in women and men's participation in these communities - differences which may further exacerbate existing gender gaps in the ICT workforce writ large.
Research questions¶
To understand the extent and nature of the gender gap in online developer communities, we ask the following research questions:
- How do male and female developers differ in their participation in online developer communities?
- How do male and female developers differ in their perceptions of belonging and kinship in online developer communities?
- How do male and female developers in online communities differ in their employment and prior experience with coding?
1. Setup¶
These lines hide the code blocks with a toggle button to display them.
These lines import the necessary package dependencies.
In [2]:
import pandas as pd
import numpy as np
import os
import csv
import matplotlib.pyplot as plt
import pandas.tools.plotting
import seaborn as sns
import matplotlib
%matplotlib inline
# Install "fuzzywuzzy" and "python-Levenshtein" using pip install
from fuzzywuzzy import fuzz
from fuzzywuzzy import process
plt.style.use('seaborn')
import plotly.offline as offline
offline.init_notebook_mode(True)
import plotly.graph_objs as go
#plt.style.use('ggplot')
pd.options.display.max_rows =300
plt.rc('ytick',labelsize=12)
plt.rc('ytick',labelsize=12)
plt.rc('axes',labelsize=12)
These lines define the relative paths to folders for data and outputs.
In [3]:
## Define folder paths
output_folder = "analysis-output"
if not os.path.exists(output_folder):
os.makedirs(output_folder)
save_images = output_folder + "\plots"
save_maps = output_folder + "\maps"
if not os.path.exists(save_images):
os.makedirs(save_images)
if not os.path.exists(save_maps):
os.makedirs(save_maps)
Make sure all the survey data is saved to your data folder.
Make sure the country/region mapping csv (found here) is saved there as well.
1.0. Load data¶
In [28]:
data_folder = os.path.abspath('..\\data')
## Load StackOverflow data
print(data_folder)
values = pd.read_csv('{0}\survey_results_public.csv'.format(data_folder), na_values=['#NULL!', 'nan'], low_memory=False)
1.1. Preview data¶
Preview column headers and sample values. Uncomment the print statements by deleting the # in front of the word print to see these values previewed.
In [9]:
## Print number of respondents
print("Total respondents:", len(values), "\n")
## Print example values for columns
#print(values.head())
## Print list of column names
#print(values.columns.ravel())
## Print all columns
#values[0:3].transpose()
1.2. Convert gender response data to binary (male/female) responses¶
Here, we use responses from survey respondents who selected either Male or Female.
Note: This is a simplification. Future analyses should investigate the nuance of response patterns for non-binary and transgender respondents.
In [31]:
## StackOverflow
values = values.fillna('Not provided')
print("StackOverflow: \n")
print(values['Gender'].value_counts(), "\n")
values.loc[(values["Gender"].str.contains('Non-binary')), "Gender"]='Non-binary'
values.loc[(values["Gender"].str.contains('Female;Male')), "Gender"]='Non-binary'
values.loc[(values["Gender"].str.contains('Transgender')), "Gender"]='Transgender'
values.loc[(values["Gender"] == ('Transgender')) | (values["Gender"] == ('Non-binary')), "Gender"]='Non-binary or Transgender'
so_female = values[values['Gender'] == "Female"]
so_male = values[values['Gender'] == "Male"]
so_gender = values[(values['Gender'] == "Male") | (values["Gender"] == "Female")]
print(so_gender["Gender"].value_counts(normalize=True)*100)
# Should we include everyone who checked Female as Female, and everyone who checked Male as Male? How would we prevent double-counting for those who checked both?
1.3. Clean country names¶
In [32]:
# Define function to check input names against ISO names
country_name_folder = os.path.abspath('..\\country-cleaning\data')
print(country_name_folder)
# Load Country-Region data
country_mapping = pd.read_csv('{0}\Country-Region-Mapping.csv'.format(country_name_folder),encoding='latin-1')
names_array=[]
ratio_array=[]
def match_names(input_names,correct_names):
for row in input_names:
x=process.extractOne(row, correct_names)
names_array.append(x[0])
ratio_array.append(x[1])
return names_array,ratio_array
Run this block first to define the name of your country variable in your source data
In [33]:
# Change this to be the name of your country name column:
country_column_name = "Country"
Run this block first to store the list of unique country names from your source data set.
In [34]:
# Create a list of unique country names from source dataset
source_names=values[country_column_name].dropna().unique()
# Create a list of country names from ISO dataset
correct_names=country_mapping['name'].values
Run this block to create a list of the name matches, with the ratio of their match (i.e. how "close" a match they are. Higher is a closer match).
In [35]:
name_match, ratio_match = match_names(source_names,correct_names)
# Check to make sure the list of unique names in the name match is the same as the list of unique countries in the source list.
# That is, there should be at least one matched name from ISO for each country name in your source data.
if len(name_match) != len(source_names):
print(len(name_match))
print(len(source_names))
print("Error: Different numbers of unique country names. Re-run code and inspect the .csv's if you continue to get this error.")
else:
print("You have", len(name_match), "country names.")
In [36]:
print(name_match, source_names)
Run this block to generate a .csv file with the source names (from your data), the ISO name it was matched with, and the % match.
Open this file and sort by % match to fix individual errors in country names (e.g. erroneously matching South Korea with Democratic People's Republic of Korea).
In [37]:
# Write an output .csv file with the source names, ISO names, and percent match between them
# Define name of output file:
output_file = "country_names_for_checking.csv"
with open('{0}\\{1}'.format(data_folder, output_file), "w") as csv_file:
writer = csv.writer(csv_file, delimiter=',')
writer.writerow(['Source Names', 'ISO Names', '% Match'])
for i in range(0, len(source_names)):
writer.writerow([source_names[i],name_match[i],ratio_match[i]])
Read fixed country names and store those in your source data frame:
In [38]:
# Read that fixed .csv and show the comparison to original source names
fixed = pd.read_csv("{0}\\{1}".format(data_folder, output_file),header=None,encoding="ISO-8859-1")
d = {'ISO Country Name':fixed[1][1:],'Original Country Name':source_names}
cross_ref =pd.DataFrame(d)
cross_ref.head()
Out[38]:
Replace survey dataset country names with correct country names.
In [39]:
# Creates a new column with the value for each row
# Define name of new column
new_country_column_name = "Country_Name"
values = values.join(cross_ref.set_index('Original Country Name'), on=country_column_name)
values = values.rename(index=str, columns={"ISO Country Name": new_country_column_name})
# Join ISO labels to source data by country name
values = values.join(country_mapping.set_index('name'), on=new_country_column_name)
# Print the number of items for each country in your data
print(values[new_country_column_name].value_counts())
Find gender ratio¶
In [41]:
# Only use countries with > 100 respondents
threshold = 100
counts = values["Country_Name"].value_counts()
country_min = values.loc[values["Country_Name"].isin(counts[counts > threshold].index), :]
# Calculate female/male ratio
female = country_min[country_min['Gender'] == 'Female']['Country_Name'].value_counts()
male = country_min[country_min['Gender'] == 'Male']['Country_Name'].value_counts()
female_share = female/(female+male).dropna()
#print(female_share)
1.2. Country Ranking by Total Respondents¶
Total number of countries¶
In [42]:
len(values["Country_Name"].value_counts())
Out[42]:
Number of countries with more than 100 respondents¶
In [43]:
len(country_min["Country_Name"].value_counts())
Out[43]:
In [44]:
country_total = country_min['Country_Name'].value_counts().plot(kind='bar',figsize=(12,10), title="StackOverflow: Total # Respondents")
fig = country_total.get_figure()
name = "SO_countries_total-min100_barplot.jpg"
save_location = save_images + "\\" + name
fig.savefig(save_location)
1.3. Country Ranking by Percent Female¶
In [45]:
plt.xticks(rotation=90)
female_share_countries = female_share.sort_values().plot(kind='bar',figsize=(15,10), title="StackOverflow: % Female Respondents")
fig = female_share_countries.get_figure()
name = "SO_countries_female-share_barplot.jpg"
save_location = save_images + "\\" + name
fig.savefig(save_location)
1.4. Regional Total Respondents¶
In [46]:
# Regional Count
#print(values['Country_Name'].value_counts())
print(values['region'].value_counts())
sns.set(font_scale=1)
count = sns.countplot(x='region', data=values)
count.set_xlabel('Region')
count.set_ylabel('Count')
count.set_title("StackOverflow: Regional Total # of Respondents")
fig = count.get_figure()
name = "SO_regional_total_barplot.jpg"
save_location = save_images + "\\" + name
fig.savefig(save_location)
print("")
In [ ]:
#### TODO: Region-level Female Percent
# print(female_share['region'].value_counts())
# sns.set(font_scale=1)
# count = sns.countplot(x='region', data=female_share)
# count.set_xlabel('Region')
# count.set_ylabel('Count')
# print("")
1.5. Map of Total # of Respondents¶
In [51]:
### Map of Total # of Respondents
country = values['Country_Name'].value_counts()
# plotly
data = [dict(
type ='choropleth',
locations = country.index,
locationmode ='country names',
z = country.values,
text = 'female respondents',
colorscale=[[0,"rgb(5, 10, 172)"],[0.35,"rgb(40, 60, 190)"],[0.5,"rgb(70, 100, 245)"],\
[0.6,"rgb(90, 120, 245)"],[0.7,"rgb(106, 137, 247)"],[1,"rgb(220, 220, 220)"]],
reversescale=True,
marker=dict(line=dict(color='rgb(180,180,180)', width=0.5)),
colorbar = dict(title ='Response count')
)]
layout = dict(title = 'StackOverflow: Total Respondents by Country',
geo = dict( showframe= False,
showcoastlines=True,
projection = dict(type = 'Mercator')))
fig = dict(data=data, layout=layout)
name = "SO_female_total"
save_location = save_maps + "\\" + name
offline.plot(fig, filename= save_location, image_filename=name,image='png')
print("Open this map: SO_female_total.png")
1.6. Map of Percent Female Respondents by Country (above 100 respondents)¶
In [52]:
### Map of Percent Female Respondents by Country (above 100 respondents)
# plotly
data = [dict(
type ='choropleth',
locations = female_share.index,
locationmode ='country names',
z = female_share.values,
text = '% female',
colorscale=[[0,"rgb(5, 10, 172)"],[0.35,"rgb(40, 60, 190)"],[0.5,"rgb(70, 100, 245)"],\
[0.6,"rgb(90, 120, 245)"],[0.7,"rgb(106, 137, 247)"],[1,"rgb(220, 220, 220)"]],
reversescale=True,
marker=dict(line=dict(color='rgb(180,180,180)', width=0.5)),
colorbar = dict(title ='Response count')
)]
layout = dict(title = 'StackOverflow: Percent Female Respondents by Country',
geo = dict( showframe= False,
showcoastlines=True,
projection = dict(type = 'Mercator')))
fig = dict(data=data, layout=layout)
name = "SO_female_percent"
save_location = save_maps + "\\" + name
offline.plot(fig, filename= save_location, image_filename=name,image='png')
print("Open this map: SO_female_percent.png")
2. Age and Years of Coding¶
In [53]:
# Function for drawing heatmaps
def draw_heatmap(column1, column2, title=None, annot=True, ax=None, size=(10, 10), data=values):
cross = pd.crosstab(data[column1], data[column2], normalize='index')
if ax is None:
f, ax = plt.subplots(figsize=size)
sns.heatmap(cross, cmap='Reds', fmt='g', annot=annot, ax=ax)
if title is not None:
ax.set_title(title)
2.1. Age and Years of Coding, by Gender¶
In [54]:
# Age and Age Started Coding
fig, ax = plt.subplots(nrows=2, ncols=2, figsize=(15,10))
sns.set()
draw_heatmap('Age', 'Gender', ax=ax[0][0], annot=True, data=values)
ax[0][0].set_xlabel("Gender")
ax[0][0].set_ylabel("Age")
draw_heatmap('YearsCoding', 'Gender', ax=ax[0][1], annot=True, data=values)
ax[0][1].set_xlabel("Gender")
ax[0][1].set_ylabel("YearsCoding")
draw_heatmap('YearsCodingProf', 'Gender', ax=ax[1][1], annot=True, data=values)
ax[1][1].set_xlabel("Gender")
ax[1][1].set_ylabel("YearsCodingProf")
fig.suptitle("StackOverflow: Age and Years of Coding, by Gender")
fig.delaxes(ax[1][0])
fig.tight_layout()
plt.subplots_adjust(top=0.95)
name = "SO_age_years-coding_gender.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
In [55]:
## Create male/female dataframes, for heatmapping
women = values[values["Gender"]=="Female"]
men = values[values["Gender"]=="Male"]
2.2. Combined Age and Years of Coding, by Gender¶
In [56]:
# Age and Age Started Coding, by Gender
fig, ax = plt.subplots(nrows=2, ncols=2, figsize=(20,17))
fig.suptitle("StackOverflow: Age and Years of Coding, by Gender")
sns.set()
draw_heatmap('YearsCoding', 'Age', ax=ax[0][0], annot=True, data=women)
ax[0][0].set_xlabel("Age")
ax[0][0].set_ylabel("YearsCoding")
ax[0][0].set_title('Women')
draw_heatmap('YearsCoding', 'Age', ax=ax[0][1], annot=True, data=men)
ax[0][1].set_xlabel("Age")
ax[0][1].set_ylabel("YearsCoding")
ax[0][1].set_title('Men')
draw_heatmap('YearsCodingProf', 'Age', ax=ax[1][0], annot=True, data=women)
ax[1][0].set_xlabel("Age")
ax[1][0].set_ylabel("YearsCodingProf")
ax[1][0].set_title('Women')
draw_heatmap('YearsCodingProf', 'Age', ax=ax[1][1], annot=True, data=men)
ax[1][1].set_xlabel("Age")
ax[1][1].set_ylabel("YearsCodingProf")
ax[1][1].set_title('Men')
fig.tight_layout()
plt.subplots_adjust(top=0.95)
name = "SO_age_years-coding_gender_combined.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
3. Education¶
In [57]:
# Create new Dataframes for each region
asia = values[values["region"] == "Asia"]
americas = values[values["region"] == "Americas"]
europe = values[values["region"] == "Europe"]
africa = values[values["region"] == "Africa"]
oceania = values[values["region"] == "Oceania"]
In [58]:
# Create new Dataframes for each Country case study
indonesia = values[values["Country_Name"] == "Indonesia"]
argentina = values[values["Country_Name"] == "Argentina"]
south_africa = values[values["Country_Name"] == "South Africa"]
pakistan = values[values["Country_Name"] == "Pakistan"]
rwanda = values[values["Country_Name"] == "Rwanda"]
3.1. Education, by Gender¶
In [59]:
# Formal Education by Gender
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(20,5))
sns.set()
draw_heatmap('FormalEducation', 'Gender', ax=ax[0], annot=True, data=values)
ax[0].set_xlabel("Gender")
ax[0].set_ylabel("FormalEducation")
fig.delaxes(ax[1])
fig.suptitle("StackOverflow: Education, by Gender")
name = "SO_education_gender.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
3.2. Education, by Gender and Region¶
In [60]:
# Formal Education, by Gender and Region
fig, ax = plt.subplots(nrows=3, ncols=2, figsize=(20,12))
sns.set()
draw_heatmap('FormalEducation', 'Gender', ax=ax[0][1], annot=True, data=africa)
ax[0][1].set_title('Africa')
sns.set()
draw_heatmap('FormalEducation', 'Gender', ax=ax[1][0], annot=True, data=americas)
ax[1][0].set_title('Americas')
sns.set()
draw_heatmap('FormalEducation', 'Gender', ax=ax[1][1], annot=True, data=asia)
ax[1][1].set_title('Asia')
sns.set()
draw_heatmap('FormalEducation', 'Gender', ax=ax[2][0], annot=True, data=europe)
ax[2][0].set_title('Europe')
sns.set()
draw_heatmap('FormalEducation', 'Gender', ax=ax[2][1], annot=True, data=oceania)
ax[2][1].set_title('Oceania')
fig.delaxes(ax[0][0])
fig.suptitle("StackOverflow: Education, by Gender and Region")
fig.tight_layout()
name = "SO_education_gender_region.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
Total respondents for 4 Countries¶
In [61]:
print("Argentina:", len(argentina), "\nIndonesia:", len(indonesia), "\nSouth Africa:", len(south_africa), "\nPakistan", len(pakistan), "\nRwanda", len(rwanda))
3.3. Education, by Gender, for 4 Countries¶
In [62]:
# Formal Education, for 4 countries
f,ax = plt.subplots(2,2, figsize=(16,10))
question = 'FormalEducation'
sns.set()
draw_heatmap(question, 'Gender', ax=ax[0][0], annot=True, data=argentina)
ax[0][0].set_title('Argentina')
sns.set()
draw_heatmap(question, 'Gender', ax=ax[0][1], annot=True, data=indonesia)
ax[0][1].set_title('Indonesia')
sns.set()
draw_heatmap(question, 'Gender', ax=ax[1][0], annot=True, data=south_africa)
ax[1][0].set_title('South Africa')
sns.set()
draw_heatmap(question, 'Gender', ax=ax[1][1], annot=True, data=pakistan)
ax[1][1].set_title('Pakistan')
plt.suptitle("StackOverflow: Education by Gender, for 4 Countries")
plt.subplots_adjust(hspace=0.5, wspace=1.6)
name = "SO_education_gender_4countries.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
4. Employment¶
4.1. Employment, by Gender¶
In [63]:
# Employment by Gender
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(20,5))
sns.set()
draw_heatmap('Employment', 'Gender', ax=ax[0], annot=True, data=values)
ax[0].set_xlabel("Gender")
ax[0].set_ylabel("Employment")
fig.suptitle("StackOverflow: Employment by Gender")
fig.delaxes(ax[1])
name = "SO_employment_gender.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
4.2. Employment, by Gender and Region¶
In [64]:
# Employment, by Gender and Region
fig, ax = plt.subplots(nrows=3, ncols=2, figsize=(20,12))
sns.set()
draw_heatmap('Employment', 'Gender', ax=ax[0][1], annot=True, data=africa)
ax[0][1].set_title('Africa')
sns.set()
draw_heatmap('Employment', 'Gender', ax=ax[1][0], annot=True, data=americas)
ax[1][0].set_title('Americas')
sns.set()
draw_heatmap('Employment', 'Gender', ax=ax[1][1], annot=True, data=asia)
ax[1][1].set_title('Asia')
sns.set()
draw_heatmap('Employment', 'Gender', ax=ax[2][0], annot=True, data=europe)
ax[2][0].set_title('Europe')
sns.set()
draw_heatmap('Employment', 'Gender', ax=ax[2][1], annot=True, data=oceania)
ax[2][1].set_title('Oceania')
fig.delaxes(ax[0][0])
fig.tight_layout()
plt.suptitle("StackOverflow: Employment by Gender, Region")
name = "SO_employment_gender_region.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
4.3. Employment, by Gender, for 4 Countries¶
In [65]:
#### Employment, for 4 countries
f,ax = plt.subplots(2,2, figsize=(16,10))
question = 'Employment'
sns.set()
draw_heatmap(question, 'Gender', ax=ax[0][0], annot=True, data=argentina)
ax[0][0].set_title('Argentina')
sns.set()
draw_heatmap(question, 'Gender', ax=ax[0][1], annot=True, data=indonesia)
ax[0][1].set_title('Indonesia')
sns.set()
draw_heatmap(question, 'Gender', ax=ax[1][0], annot=True, data=south_africa)
ax[1][0].set_title('South Africa')
sns.set()
draw_heatmap(question, 'Gender', ax=ax[1][1], annot=True, data=pakistan)
ax[1][1].set_title('Pakistan')
plt.suptitle("StackOverflow: Employment by Gender, for 4 Countries")
plt.subplots_adjust(hspace=0.5, wspace=1.6)
name = "SO_employment_gender_4countries.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
4.4. Employment, by Company Size¶
In [66]:
print(values["CompanySize"].value_counts())
In [67]:
## Subset only those employed at a company
employed = values[(values["Employment"] == "Employed full-time") | (values["Employment"] == "Employed part-time")]
print(employed["Employment"].value_counts())
female_employed = employed[employed["Gender"] == "Female"]
male_employed = employed[employed["Gender"] == "Male"]
In [68]:
# Employment by Company Size
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(25,3))
sns.set()
draw_heatmap('Employment', 'CompanySize', ax=ax[0], annot=True, data=employed)
ax[0].set_xlabel("CompanySize")
ax[0].set_ylabel("Employment")
fig.suptitle("StackOverflow: Employment by CompanySize")
fig.delaxes(ax[1])
name = "SO_employment_company_size.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
4.5. Employment, by Company Size and Gender¶
In [69]:
# Employment by Company Size, Gender
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(25,3))
sns.set()
draw_heatmap('Employment', 'CompanySize', ax=ax[0], annot=True, data=female_employed)
ax[0].set_xlabel("CompanySize")
ax[0].set_ylabel("Employment")
ax[0].set_title('Female Employment')
draw_heatmap('Employment', 'CompanySize', ax=ax[1], annot=True, data=male_employed)
ax[1].set_xlabel("CompanySize")
ax[1].set_ylabel("Employment")
ax[1].set_title('Male Employment')
fig.suptitle("StackOverflow: Employment by CompanySize and Gender")
name = "SO_employment_company_size.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
5. Company Size¶
5.1. Company Size, by Gender¶
In [70]:
# Company Size by Gender
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(20,5))
sns.set()
draw_heatmap('CompanySize', 'Gender', ax=ax[0], annot=True, data=values)
ax[0].set_xlabel("Gender")
ax[0].set_ylabel("CompanySize")
fig.suptitle("StackOverflow: CompanySize by Gender")
fig.delaxes(ax[1])
name = "SO_company_size_gender.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
5.2. Company Size, by Region¶
In [71]:
# Company Size by Region
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(20,5))
sns.set()
draw_heatmap('CompanySize', 'region', ax=ax[0], annot=True, data=employed)
ax[0].set_xlabel("Region")
ax[0].set_ylabel("CompanySize")
fig.suptitle("StackOverflow: CompanySize by Region")
fig.delaxes(ax[1])
name = "SO_company_size_region.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
5.3. Company Size, by Gender and Region¶
In [72]:
# Company Size, by Gender and Region
fig, ax = plt.subplots(nrows=3, ncols=2, figsize=(20,12))
sns.set()
draw_heatmap('CompanySize', 'Gender', ax=ax[0][1], annot=True, data=africa)
ax[0][1].set_title('Africa')
sns.set()
draw_heatmap('CompanySize', 'Gender', ax=ax[1][0], annot=True, data=americas)
ax[1][0].set_title('Americas')
sns.set()
draw_heatmap('CompanySize', 'Gender', ax=ax[1][1], annot=True, data=asia)
ax[1][1].set_title('Asia')
sns.set()
draw_heatmap('CompanySize', 'Gender', ax=ax[2][0], annot=True, data=europe)
ax[2][0].set_title('Europe')
sns.set()
draw_heatmap('CompanySize', 'Gender', ax=ax[2][1], annot=True, data=oceania)
ax[2][1].set_title('Oceania')
fig.delaxes(ax[0][0])
fig.tight_layout()
plt.suptitle("StackOverflow: CompanySize by Gender, Region")
name = "SO_company_size_gender_region.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
5.4. Company Size, by Gender and Region (Alternative view)¶
In [73]:
# Company Size, by Gender, Region
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(25,6))
sns.set()
draw_heatmap('CompanySize', 'region', ax=ax[0], annot=True, data=male_employed)
ax[0].set_xlabel("CompanySize")
ax[0].set_ylabel("Region")
ax[0].set_title('Men')
draw_heatmap('CompanySize', 'region', ax=ax[1], annot=True, data=female_employed)
ax[1].set_xlabel("CompanySize")
ax[1].set_ylabel("Region")
ax[1].set_title('Women')
fig.suptitle("StackOverflow: CompanySize by Region and Gender")
name = "SO_company_size_gender_region.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
6. Computer Science (and related) Degrees¶
In [74]:
print(values["UndergradMajor"].value_counts())
In [75]:
## Create male/female dataframes for CS and non-CS graduates, for heatmapping
CS_overall = values[(values["UndergradMajor"]=="Computer science, computer engineering, or software engineering") | (values["UndergradMajor"]== "Information systems, information technology, or system administration") | (values["UndergradMajor"] == "Web development or web design")]
non_CS_overall = values[(values["UndergradMajor"]!="Computer science, computer engineering, or software engineering") & (values["UndergradMajor"]!= "Information systems, information technology, or system administration") & (values["UndergradMajor"] != "Web development or web design")]
CS_women = women[(women["UndergradMajor"]=="Computer science, computer engineering, or software engineering") | (women["UndergradMajor"]== "Information systems, information technology, or system administration") | (women["UndergradMajor"] == "Web development or web design")]
non_CS_women = women[(women["UndergradMajor"]!="Computer science, computer engineering, or software engineering") & (women["UndergradMajor"]!= "Information systems, information technology, or system administration") & (women["UndergradMajor"] != "Web development or web design")]
CS_men = men[(men["UndergradMajor"]=="Computer science, computer engineering, or software engineering") | (men["UndergradMajor"]== "Information systems, information technology, or system administration") | (men["UndergradMajor"] == "Web development or web design")]
non_CS_men = men[(men["UndergradMajor"]!="Computer science, computer engineering, or software engineering") & (men["UndergradMajor"]!= "Information systems, information technology, or system administration") & (men["UndergradMajor"] != "Web development or web design")]
6.1. CS Degrees, by Gender¶
In [76]:
# CS Degrees, by Gender
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(25,6))
sns.set()
draw_heatmap('UndergradMajor', 'Gender', ax=ax[0], annot=True, data=values)
ax[0].set_xlabel("Gender")
ax[0].set_ylabel("UndergradMajor")
ax[0].set_title('Men')
fig.suptitle("StackOverflow: CS Degrees, by Gender")
fig.delaxes(ax[1])
name = "SO_degree_gender.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
6.2. CS Degrees, by Gender and Region¶
In [77]:
# CS Degrees, by Gender and Region
fig, ax = plt.subplots(nrows=3, ncols=2, figsize=(20,12))
sns.set()
draw_heatmap('UndergradMajor', 'Gender', ax=ax[0][1], annot=True, data=africa)
ax[0][1].set_title('Africa')
sns.set()
draw_heatmap('UndergradMajor', 'Gender', ax=ax[1][0], annot=True, data=americas)
ax[1][0].set_title('Americas')
sns.set()
draw_heatmap('UndergradMajor', 'Gender', ax=ax[1][1], annot=True, data=asia)
ax[1][1].set_title('Asia')
sns.set()
draw_heatmap('UndergradMajor', 'Gender', ax=ax[2][0], annot=True, data=europe)
ax[2][0].set_title('Europe')
sns.set()
draw_heatmap('UndergradMajor', 'Gender', ax=ax[2][1], annot=True, data=oceania)
ax[2][1].set_title('Oceania')
fig.delaxes(ax[0][0])
fig.tight_layout()
plt.suptitle("StackOverflow: CS Degrees by Gender, Region")
name = "SO_degree_gender_region.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
6.3. CS Degrees, by Employment and Gender¶
In [78]:
# CS Degrees, by Employment and Gender
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(25,6))
sns.set()
draw_heatmap('Employment', 'Gender', ax=ax[0], annot=True, data=CS_overall)
ax[0].set_xlabel("Employment")
ax[0].set_ylabel("Gender")
ax[0].set_title('CS Majors')
draw_heatmap('Employment', 'Gender', ax=ax[1], annot=True, data=non_CS_overall)
ax[1].set_xlabel("Employment")
ax[1].set_ylabel("Gender")
ax[1].set_title('Non-CS Majors')
fig.suptitle("StackOverflow: CS Degrees, by Employment and Gender")
name = "SO_degree_gender_employment.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
6.4. Women with CS Degrees, by Employment and Region¶
In [79]:
# Women with CS Degrees, by Employment and Region
fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(25,6))
sns.set()
draw_heatmap('Employment', 'region', ax=ax[0], annot=True, data=CS_women)
ax[0].set_xlabel("Employment")
ax[0].set_ylabel("Region")
ax[0].set_title('CS Major Women')
draw_heatmap('Employment', 'region', ax=ax[1], annot=True, data=non_CS_women)
ax[1].set_xlabel("Employment")
ax[1].set_ylabel("Region")
ax[1].set_title('Non-CS Major Women')
fig.suptitle("StackOverflow: Women with CS Degrees, by Employment and Region")
name = "SO_degree_employment_region+_female.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
7. Job Level¶
7.1. Developers, by Gender¶
In [80]:
developers = values[values['DevType'].str.contains("developer")]
print(developers["Gender"].value_counts())
In [81]:
## Barchart of Gender of Developers
sns.set(font_scale=1)
count = sns.countplot(x='Gender', data=developers)
count.set_xlabel('Gender')
count.set_ylabel('Count')
count.set_title("StackOverflow: Developer Gender")
fig = count.get_figure()
name = "SO_devtype_dev_gender_barplot.jpg"
save_location = save_images + "\\" + name
fig.savefig(save_location)
print("")
7.2. Managers, by Gender¶
In [82]:
managers = values[values['DevType'].str.contains("manager")]
print(managers["Gender"].value_counts())
In [83]:
## Barchart of Gender of Managers
sns.set(font_scale=1)
count = sns.countplot(x='Gender', data=managers)
count.set_xlabel('Gender')
count.set_ylabel('Count')
count.set_title("StackOverflow: Manager Gender")
fig = count.get_figure()
name = "SO_devtype_manager_gender_barplot.jpg"
save_location = save_images + "\\" + name
fig.savefig(save_location)
print("")
7.3. Executives, by Gender¶
In [84]:
executives = values[values['DevType'].str.contains("executive")]
print(executives["Gender"].value_counts())
In [85]:
## Barchart of Gender of C-Suite Executives
sns.set(font_scale=1)
count = sns.countplot(x='Gender', data=executives)
count.set_xlabel('Gender')
count.set_ylabel('Count')
count.set_title("StackOverflow: C-Suite Executive Gender")
name = "SO_devtype_executive_gender_barplot.jpg"
save_location = save_images + "\\" + name
fig.savefig(save_location)
print("")
7.4. Job Types, by Gender and Region¶
In [86]:
# Developers by Gender, Region
fig, ax = plt.subplots(nrows=2, ncols=2, figsize=(20,5))
sns.set()
draw_heatmap('Gender', 'region', ax=ax[0][0], annot=True, data=developers)
ax[0][0].set_xlabel("Region")
ax[0][0].set_ylabel("Gender")
ax[0][0].set_title("Developers")
draw_heatmap('Gender', 'region', ax=ax[0][1], annot=True, data=managers)
ax[0][1].set_xlabel("Region")
ax[0][1].set_ylabel("Gender")
ax[0][1].set_title("Managers")
draw_heatmap('Gender', 'region', ax=ax[1][0], annot=True, data=executives)
ax[1][0].set_xlabel("Region")
ax[1][0].set_ylabel("Gender")
ax[1][0].set_title("C-Suite Executives")
fig.delaxes(ax[1][1])
plt.suptitle("StackOverflow: Job Role by Gender, Region")
fig.tight_layout()
name = "SO_devtype_gender_region.jpg"
save_location = save_images + "\\" + name
plt.savefig(save_location)
